kr_lp_pgnet/configs/kr_lp_pgnet.yml

Global:
  use_gpu: True
  epoch_num: 100
  log_smooth_window: 20
  print_batch_step: 10
  save_model_dir: ./output/kr_lp_pgnet/
  save_epoch_step: 10
  eval_batch_step: [0, 1000]
  cal_metric_during_train: False
  pretrained_model:
  checkpoints:
  save_inference_dir:
  use_visualdl: False
  use_wandb: True
  infer_img:
  infer_visual_type: CN
  valid_set: partvgg
  save_res_path: ./output/kr_lp_pgnet/predicts.txt
  character_dict_path: ppocr/utils/kr_lp_dict.txt
  character_type: CN
  max_text_length: 10
  max_text_nums: 5
  tcl_len: 64

wandb:
  project: kr_lp_pgnet
  name: step1-pretrain

Architecture:
  model_type: e2e
  algorithm: PGNet
  Transform:
  Backbone:
    name: ResNet
    layers: 50
  Neck:
    name: PGFPN
  Head:
    name: PGHead
    character_dict_path: ppocr/utils/kr_lp_dict.txt

Loss:
  name: PGLoss
  tcl_bs: 64
  max_text_length: 10
  max_text_nums: 5
  pad_num: 67

Optimizer:
  name: Adam
  beta1: 0.9
  beta2: 0.999
  lr:
    name: Cosine
    learning_rate: 0.001
    warmup_epoch: 5
  regularizer:
    name: 'L2'
    factor: 0.0001

PostProcess:
  name: PGPostProcess
  score_thresh: 0.5
  mode: fast
  point_gather_mode: align

Metric:
  name: E2EMetric
  mode: A
  gt_mat_dir: ./train_data/kr_lp_synth/gt
  character_dict_path: ppocr/utils/kr_lp_dict.txt
  main_indicator: f_score_e2e

Train:
  dataset:
    name: PGDataSet
    data_dir: ./train_data/kr_lp_synth/train
    label_file_list: [./train_data/kr_lp_synth/train/train.txt]
    ratio_list: [1.0]
    transforms:
      - DecodeImage:
          img_mode: BGR
          channel_first: False
      - E2ELabelEncodeTrain:
      - PGProcessTrain:
          batch_size: 16
          use_resize: True
          use_random_crop: False
          min_crop_size: 24
          min_text_size: 4
          max_text_size: 512
          point_gather_mode: align
      - KeepKeys:
          keep_keys: ['images', 'tcl_maps', 'tcl_label_maps', 'border_maps', 'direction_maps', 'training_masks', 'label_list', 'pos_list', 'pos_mask']
  loader:
    shuffle: True
    drop_last: True
    batch_size_per_card: 16
    num_workers: 8

Eval:
  dataset:
    name: PGDataSet
    data_dir: ./train_data/kr_lp_synth/test
    label_file_list: [./train_data/kr_lp_synth/test/test.txt]
    transforms:
      - DecodeImage:
          img_mode: BGR
          channel_first: False
      - E2ELabelEncodeTest:
      - E2EResizeForTest:
          max_side_len: 768
      - NormalizeImage:
          scale: 1./255.
          mean: [0.485, 0.456, 0.406]
          std: [0.229, 0.224, 0.225]
          order: 'hwc'
      - ToCHWImage:
      - KeepKeys:
          keep_keys: ['image', 'shape', 'polys', 'texts', 'ignore_tags', 'img_id']
  loader:
    shuffle: False
    drop_last: False
    batch_size_per_card: 1
    num_workers: 2